Agencies saving one-fifth of IT budget through virtualization, survey finds
By
Joseph Marks
More than 80 percent of federal information technology leaders say their agencies have implemented some manner of server virtualization and, overall, the government is saving nearly 20 percent of its IT budget through virtualization, according to a recent survey.
The majority of federal agencies also plan to virtualize some desktop applications, but less than 10 percent expect to do so with all desktop components for all employees, according to the survey conducted by MeriTalk, a government IT industry group. The survey was sponsored by Microsoft and NetApp, a computer storage and software provider.
Virtualization essentially allows users to complete computer tasks with Web-based tools rather than physical ones. In the case of virtualized servers it means those servers are able to use deceptively simple hardware to do complex tasks. For virtualized desktops, workers can access Web-based versions of their desktop tools from a home computer, tablet or smartphone.
Server virtualization is a major component of the government's transition to cloud computing storage, which IT leaders say will save about $5 billion annually by 2015. It's also critical to a related plan to shutter or consolidate about 40 percent of the government data centers during the next several years.
Cloud-based computing systems will allow for virtual desktops, which telework advocates say will help employees be as productive at home as they are in the office.
About 37 percent of the federal IT workload today is done on virtualized servers and technology managers expect that to grow to about 63 percent by 2015, the survey found.
Among the greatest challenges to virtualization in the federal government are legacy systems that can't easily be worked into a virtualized network and a lack of available funds, respondents said.
The MeriTalk survey included about 300 federal, state and local government IT leaders. It had a margin of error of about 6 percent. Margins of error are notoriously difficult to calculate in federal IT surveys because the pool of respondents is small and responses can vary widely among different types of agencies.